Prior studies have shown increased insulin (INS) action both in in vivo and in vitro after INS therapy in patients with NIDDM. To determine whether this increase was due to INS therapy itself or causally related to improve glycemia, we studied 9 obese Pima Indians with NIDDM (wt=99.9 plus/minus 8.2 kg; fat = 39 plus/minus 2% by underwater weighing) during weight maintenance periods. Following 5 weeks of weight loss (6.7 kg plus/minus 1.3 p. less than .001), fasting glucose fell from 251 plus/minus 10 to 148 plus/minus 15 mg/dl (p less than .0001), and 3 hr. postprandial IHS area increased by 42 plus/minus 11% (p. less than .02). Hepatic glucose production measured by 1-C13-glucose and corrected for glucose recycling fell by 38 plus/minus 3% post weight loss (4.0 plus/minus .2 to 2.8 plus/minus .2 mg,kg FFM min, p less than .0002). Glucose disposal during hyperglycemic (313 plus/minus 2mg/dl) clamps with somatostatin infusion (250 mug/hr) at 3 INS (12, 100,, 2600 muU/ml increased 20-30% after weight loss (4.8 plus/minus .2 to 5.2 plus/minus 4, 4.3 + .2 to 5.3 plus/minus.4, 11.4 + .6 to 15.6 plus/minus 1.4 mg/kg FFM min, NS, p less than .01, respectively, for the 3 INS steps). In the adipocytes there were no significant changes in basal or maximum INS stimulated glucose transport, basal lipolysis, or INS binding. We conclude that weight loss in NIDDM is associated with decreases in glycemia, which may be secondary to decreases in hepatic glucose production and increases in insulin action in vivo, and no change in in vitro insulin action in adipocytes. Previously reported increases in glucose transport in adipocytes after insulin therapy in NIDDM were probably due to the insulin therapy per se and not causally related to lower glycemia.